Sustainable Computing Conference Sessions

How do you enhance and measure your data center’s energy efficiency? Industry experts from NVIDIA, Vertiv, Cadence, and Isambard AI will share real-world examples of cutting-edge solutions designed to enhance the sustainability of data center operations. They'll cover strategies for optimizing energy consumption, water efficiency, and reducing carbon footprints, highlighting the latest innovations in accelerated computing, cooling, digital twins, and sustainable supercomputing.

AI equals more data centers, and more data centers equals more energy. Curbing carbon emissions is a growing concern for data center operators. Discover how AI data centers of the future will be the most sustainable data centers, and the criticality of power and cooling infrastructure in this equation.

Important: Near capacity, highly suggest arriving early. Attendees are let in on a first-come, first-served basis.

With the growing adoption of AI, organizations are navigating a complex AI landscape shaped by unique business needs, data residency, privacy, and regulatory requirements. To address these challenges, many enterprises are shifting towards a distributed infrastructure for AI, with data and compute spanning private infrastructure and public clouds. In such a distributed environment, many enterprises require a private AI deployment that can deliver the highest level of data privacy and control over their AI operations, while also focusing on sustainability. We will explore how organizations are navigating the complex AI landscape, maintaining the highest level of data privacy and control over their AI operations, while still prioritizing sustainability.

AI is transforming industries and driving solutions to some of the world’s most pressing challenges. To support this growing demand, the infrastructure powering AI must also evolve — with a focus on sustainability and energy efficiency. In this joint session, NVIDIA and Crusoe will discuss how advanced networking solutions are enabling a new era of AI-driven innovation. Discover how Crusoe leverages NVIDIA’s accelerated networking platforms to build a sustainable AI cloud that delivers high performance while minimizing energy consumption. Learn how Crusoe’s vertically integrated infrastructure, powered by innovative energy sources and NVIDIA’s networking technology, combines enterprise-class controls, scalability, and sustainability to meet the demands of next-generation AI workloads.

"Earth to AI" brings together industry leaders to explore how AI and climate science are transforming business strategies for a sustainable future. From cutting-edge technologies to innovative funding models, this panel highlights actionable ways to drive climate resilience and economic growth.

As AI workloads grow in scale, complexity, and intensity, the demand for high-density, scalable, and rapidly deployable power and cooling infrastructure has never been greater. But challenges like power availability, permitting delays, infrastructure densification, and the need for dynamic, AI-specific solutions threaten to slow progress. Join us for a cutting-edge discussion on the next generation of AI-ready infrastructure, from power-efficient architectures to liquid cooling breakthroughs, and discover how we can future-proof AI factories for demands of tomorrow.

This panel explores the transformative potential of artificial intelligence and accelerated computing in deciphering the complex drivers of urban climate change, and how it can help mitigate it. As cities, home to 4.4 billion people, grapple with increasing environmental challenges, including extreme heat waves driven by a combination of climate change and the urban heat island effect, these cutting-edge technologies offer unprecedented capabilities to analyze vast datasets, model intricate urban systems, and generate actionable insights for policymakers and urban planners. We'll delve into how AI-powered tools are revolutionizing climate modeling and weather forecasting, enabling high-resolution simulations of urban microclimates and their interactions with the built environment. Experts will discuss the application of generative AI, digital twins, and machine learning algorithms in processing and interpreting diverse data sources, including satellite imagery, internet-of-things sensor networks, and social media feeds, to provide a comprehensive understanding of urban heat islands, air pollution patterns, and energy consumption dynamics. We'll highlight accelerated computing's role in dramatically reducing computation time and energy consumption for climate simulations and show how it enables more frequent and detailed analyses of urban climate phenomena. We'll also cover innovative projects leveraging these technologies, such as digital twins of cities and AI-driven urban planning tools that optimize for climate resilience. Panelists will discuss the challenges and opportunities in implementing AI and accelerated computing solutions, including data privacy concerns, the need for interdisciplinary collaboration, the critical role of compute infrastructure, and strategies for ensuring equitable access to these technologies. We'll finish by exploring the potential of these technologies to inform evidence-based climate adaptation strategies and contribute to sustainable development goals.

Learn from the global industry thought leaders who are shaping the power industry to meet demands of the energy transition. This lightning talk session will bring Southern California Edison, Saudi Electric, Energy Power Research Institute, and Siemens Energy together to discuss the use of generative AI to better predict demand forecasting for distributed energy resources, and to make grid resilient.

Groundbreaking hardware technologies and AI are unlocking unprecedented computational power. NVIDIA’s Ian Buck will highlight cutting-edge innovations that are revolutionizing high-performance computing while prioritizing energy conservation. Discover how these technologies are enabling industries and nations to achieve new insights. Join us to explore how supercomputing is not only advancing capabilities, but also paving the way for a greener, more efficient future in tech.

Future-proofing your AI requires a dynamic platform, accelerated infrastructure, and multicloud interconnection to scale effectively. An integrated end-to-end solution can speed up AI deployment, while accelerated infrastructure optimizes the training and development of AI models. As AI drives higher demand for data center supercomputing, enterprises will continue to seek a flexible, predictable, and sustainable approach to acquiring, deploying, and managing accelerated computing infrastructure — without the complexity of designing a data center that can accommodate high-density power and advanced liquid cooling in a sustainable way. Learn how businesses can architect a hybrid multicloud-connected AI data center that offers greater flexibility, scalability, and sustainability.

Forty percent of a typical data center's energy consumption comes from cooling, and AI is a power-hungry workload. How can more organizations benefit from high-performance, powerful AI compute with greater energy efficiency? Hear strategies and solutions that can meet the increasing compute demands of AI while addressing energy efficiency through innovative thermal management. Thermal challenges are an unavoidable outcome to rapid innovation of the GPU to serve AI workloads. Liquid-cooled solutions are being deployed into next-gen AI factories to deliver performance improvements and energy savings. Learn about Lenovo’s expertise and consistent DNA of liquid-cooled designs since 2012, enabling no-limits AI and helping meet your sustainability goals.

Our experiments show partitioning and multiplexing the GPUs with multi-process service and multi-instance GPUs allows for higher throughput and higher energy efficiency for ML and HPC workloads. We present our analysis with different generation GPUs and show more than 25% energy efficiency compared to running applications without partitioning the GPUs.

Find out Delta’s new developments in grid-to-chip power solutions to optimize energy efficiency in gigawatt-scale data centers for AI training.

The trends in computer architecture, primarily driven by AI-based applications (most recently, large language models), has led to a rapid increase in the reduced- and mixed-precision computing capabilities of GPUs. These processors demonstrate an outsized power-efficiency (FLOPS/watt) advantage over systems almost exclusively focused upon native single- and double-precision arithmetic. Thus, there is a great deal of motivation to leverage these capabilities, through the use of various mixed-precision algorithms and emulation techniques, to facilitate greater scientific computing throughput without sacrificing accuracy. We'll touch upon a number of these approaches and present real-world case studies that provide compelling evidence in support of this path to increasing the science per watt of supercomputers.

It's becoming increasingly vital to optimize for energy and power, in addition to raw performance, for environmental, financial, and logistical reasons. We'll showcase recent work to tune a range of representative HPC and AI benchmarks and applications for energy and power, where the findings and techniques are transferable to other applications. We'll show how careful tuning of the industry-standard High Performance Linpack (HPL) benchmark within the performance and energy landscape on Grace Hopper has enabled the NVIDIA-powered supercomputers to sweep the Green500 world rankings. We'll show the effectiveness of tuning processor clock frequencies for energy minimization across a range of applications, and how application-level optimizations can go even further. We'll discuss this work within the context of the latest NVIDIA technology.

Discover Pegatron's groundbreaking high-density GPU rack solution, leveraging the NVIDIA platform to revolutionize AI infrastructure. We'll explore the innovative design and performance enhancements that drive efficiency and scalability in AI applications, addressing the data center's demand for higher-density AI computing power, key factors in high-density design, and the importance of liquid cooling.